I have only seen this matter from the other perspective, am scientist, not technically a CS person but I do programming/data analytics for work. I saw...

• A colleague of mine, who teaches a statistics course, mentioned that literally every student of his uses LLMs aggressively for homework assignments. Worse, a lot of his students would just plug the assignment into an LLM, and if he quizzes/asks them for reason, they don't know why 💀

• It gets worse. So I have asked ChatGPT to make a powerpoint template once, but didn't like it so I didn't ever ask it again. However, after that exercise, I started to notice the exact template that Chat gave me for the slides to appear... in every scientific meeting I go to... among my coworkers slides... at startup/tech events... everywhere

• It gets even worse. Based on what me and my colleagues observe, my supervisor straight-up prefers slides and presentations with that distinct AI-generated style. The workloads we are given also basically more-or-less expects us to heavily use some type of agentic workflow. When I'm searching for jobs, a lot of data engineer/software job openings directly mention that they either want someone experienced in agentic/vibe coding, or want ppl who are not morally against AI

So yeah we are in strange times. I don't know what schools/unis are supposed to do... probably give them long enough time and they can come up with a solution, similar to how high-schools treat graphics calculators. At least my hopium says that, from a learning/job training perspective, someone who is ok with agentic coding but not blinded by it probably has pretty good job prospects so